Performances of Antibiotic Screening Tests in Determining the Persistence of Penicillin Residues in Ewe s Milk

2097 Journal of Food Protection, Vol. 66, No. 11, 2003, Pages 2097 2102 Copyright q, International Association for Food Protection Performances of Antibiotic Screening Tests in Determining the Persistence of Penicillin Residues in Ewe s Milk M. I. BERRUGA, 1 M. YAMAKI, 1 R. L. ALTHAUS, 2 M. P. MOLINA, 3 AND A. MOLINA 1 * 1 Departamento de Ciencia y Tecnolog a Agroforestal, ETSIA, Universidad de Castilla-La Mancha, 02071 Albacete, Spain; 2 Cátedra de Biof sica, Facultad de Ciencias Veterinarias, Universidad Nacional del Litoral, 3080 Esperanza, Argentina; and 3 Departamento de Ciencia Animal, Universidad Politécnica, 46071 Valencia, Spain MS 03-7: Received 10 January 2003/Accepted 14 May 2003 ABSTRACT Milk collected at 12-h intervals throughout 6 days from three groups of Manchega ewes (n 5 12 per group), treated intramuscularly with b-lactams (benzyl penicillin procaine, ampicillin, and amoxicillin, respectively), was evaluated for antibiotic residue persistence with four microbial inhibitor tests (BRT MRL, CH ATK P&S, Delvotest SP, and Eclipse 100ov) and one enzymatic test (Penzym 100). Antibiotic depletion time was established using a logistic regression model. A clear effect (P, 0.0001) of milking order on the response of all tests was observed with the three antibiotics, but no signi cant effects were found for milk yield. Except with Eclipse 100ov, positive tests were observed after the recommended withdrawal period of benzyl penicillin procaine ( ve milkings) from 2% (Delvotest SP) to 11% (CH ATK P&S). There were almost no positive responses beyond the withdrawal period (six milkings) of ampicillin, except for the Penzym 100 test (7%). Residues of amoxicillin were found to persist beyond the six milkings established as the withdrawal period, from 2.8% (Eclipse 100ov) to 72.4% (CH ATK P&S) of positive cases. Higher frequencies of doubtful cases were found with BRT MRL and Delvotest SP assays with the three b-lactams. Positive and doubtful results could be obtained when milk samples from individual ewes were analyzed using BRT MRL, CH ATK P&S, Delvotest SP, and Penzym 100 tests, even if farmers follow the antibiotic withdrawal periods. The Mediterranean Basin is an important producer of milk from ewes and goats. Most of the ewe s milk production from this area is used in cheese manufacture, often as raw milk. As intensi cation of milk production in small ruminants has increased of late, the use of antimicrobial substances in dairy ewes has become the usual practice in veterinary medicine to treat mastitis and other diseases. This use could have serious consequences in public health and dairy product quality; antibiotic residues in milk might provoke antibiotic resistance (18, 33) or allergies in consumers (29) and cause defects in several fermented products (24, 30), among other problems. To avoid the aforementioned risks, the European Union (EU) has passed Council Directive EEC 2377/90 (10), and subsequent amendments (almost 100), establishing the maximum antibiotic residue levels in animal food. Therefore, the withdrawal periods are calculated according to these maximum residue limits (MRLs), taking into account several factors, such as animal species, administration route, dosage, and so on (8, 19). The most important way to guarantee residue levels in milk below the established MRLs is by respecting withholding periods after antibiotic treatment. The farmer s responsibility is essential at this point to maintain the safety of any animal food produced (11). However, because there are few speci c available antibiotic treatments for dairy * Author for correspondence: Tel: 134 902 204 100 Ext 2030; Fax: 134 902 204 230; E-mail: ana.molina@uclm.es. ewes compared to those for cows, the withdrawal times have not been well established in this species (17). In order to determine the presence of antimicrobial residues in milk and to provide toxicologicaland technological safety, several rapid screening methods have been developed to test milk on the farm (16, 21, 25). These methods should be used as the rst step to detect antimicrobial residues, as is recommended by the integrated system for detection of antimicrobials and inhibitors in milk (20). Most of these screening tests have been commercially available for cow s milk. However, there are currently several screening tests for evaluating ewe s milk, e.g., BRT (4) or Penzym (3, 27), which give highly sensitive results in the detection of antimicrobial residues in milk for this species. The detection limits of some of these methods for ewe s milk are summarized in Table 1. Microbial inhibition tests have been employed to evaluate the antibiotic depletion time in dairy cattle (9, 23, 31, 34). In dairy sheep, there are previous studies about the effect of several different antibiotic treatments or administration routes (5, 26) on the response of some microbial inhibition tests. However, the behavior of several different methods throughout the depletion time has not been studied. For this reason, the aim of this study was to analyze the presence of residual antibiotics in milk after treatment of ewes with different penicillins by several recommended screening methods. An attempt has also been made to determine whether ewe s milk could show signs of residual

2098 BERRUGA ET AL. J. Food Prot., Vol. 66, No. 11 TABLE 1. Detection limits (mg/kg) of b-lactam antibiotics tested by various methods in ewe s milk Detection limit (mg/kg) b-lactam BRT Delvotest SP Eclipse 100ov Penzym 100 Amoxicillin Ampicillin Cloxacillin Penicillin G Cefadroxil Cephalosporin C Cephalexin Cefoperazone Ceftiofur Cefuroxime 6 6 51 2 230 1,330 270 92 120 69 5 3 23 1.4 63 610 68 41 59 41 Reference(s) 4 7 6, 28 3 7 4 5 86 115 107 85 5 4 33 3 43 10 16 900 120 77 penicillin even after ful lling the manufacturers recommended withdrawal period. MATERIALS AND METHODS Animals. Thirty-six healthy Manchegan-bred ewes (body weight range 55 to 80 kg, age from 3 to 7 years) from the Manchegan herd in the experimental ock of the Instituto Técnico Agronómico Provincial (Diputación de Albacete, Albacete, Spain) were selected for this study. None had a previous history of clinical mastitis, and all ewes presented somatic cell counts (SCC) of,100,000 somatic cells per ml before starting the treatment. The ewes were randomly divided into three groups, each containing 12 animals roughly equal in body weight, age, and daily milk yield. The animals received no drugs before the experiment was carried out. The ewes were housed and fed the same as the rest of the herd. Lambing took place in January and February, and mothers suckled their lambs for 35 days. After weaning, the ewes were machine-milked twice a day. The experiment took place in the fourth week of the milking period. Antibiotic treatment and sample collection. Antibiotics were administered immediately after milking was completed. Each group of 12 animals was treated with two intramuscular injections of a different antibiotic at 48-h intervals. The rst group was given 15,000 IU/kg body weight of benzyl penicillin procaine (Depocillin, Intervet, Salamanca, Spain). The second group was treated with 1 ml/10 kg body weight of a suspension of 150 mg/ml of amoxicillin trihydrate (Clamoxyl LA, P zer, Madrid, Spain). The third group was given 0.15 ml/kg body weight of 100 mg/ml of ampicillin (Am pen LA, Intervet). The products were administered according to the manufacturers instructions. After treatment, the ewes were machine-milked twice a day. The milk yield from each ewe was measured, and samples were taken for analysis. Additional milk samples were collected every 12 h for up to 12 milking sessions. The withdrawal period recommended by manufacturers was ve milkings for benzyl penicillin procaine and six milkings for amoxicillin and ampicillin. The samples (50 ml/milking session/animal) were collected in disposable plastic containers and kept at 48C until they were analyzed. Antibiotic detection tests. All milk samples were analyzed during the 24-h period after collection by simultaneously using four microbial growth inhibition assaysbrilliant black reduction MRL test (BRT MRL), Copan test microplate P&S (CH ATK P&S), Delvotest SP, and Eclipse 100ovand an enzymatic colorimetric assay (Penzym 100). In all ve tests, milk shown to be free from antimicrobials was used as negative control. This milk came from healthy animals that did not receive feed or treatments containing drugs. For the positive control, 4 mg benzyl penicillin G (Pen Na; SIGMA) per kg was added to negative milk samples. Visual interpretation was carried out independently by three trained observers and evaluated visually as negative, doubtful, or positive. BRT MRL. One hundred microliters of milk was added to ready-prepared individual BRT MRL (AIM-Analytik in Milch Produktions- und Vertriebs GmbH, Munich, Germany) wells containing Bacillus stearothermophilus var. calidolactis. After a 1-h diffusion period at 48C, the BRT MRL tests were incubated in a dry block heater at 64 6 18C for 2 h 30 min. CH ATK P&S. After addition of 100 ml of milk samples into single wells of CH ATK P&S (Chr Hansen, Hoersholm, Denmark) containing spores of B. stearothermophilusvar. calidolactis, plates were incubated at 64 6 18C for 3 h in a dry block heater. Delvotest SP. One nutrient tablet and 100 ml of each milk sample were added to individual wells of Delvotest SP (DSM Food Specialties, Delft, The Netherlands), ready-prepared, containing B. stearothermophilus var. calidolactis. Plates were incubated in a dry block heater at 64 6 18C for 2 h 45 min. Eclipse 100ov. Fifty microliters of each milk sample plus 50 ml of ov 100 solution (ZEU-Inmunotec, Zaragoza, Spain) were added to each well of Eclipse 100ov test (ZEU-Inmunotec), readyprepared, containing B. stearothermophilus var. calidolactis. Before being incubated at 64 6 18C for 2 h 30 min in a dry block heater, plates were preincubated 1 h at room temperature for milk and antibiotic diffusion. Penzym 100. The Penzym 100 test (Chr Hansen) was carried out according to the manufacturer s label inserts. The incubation period after addition of the chromogenic substrate was 22 min (27). Visual interpretation of the results by three trained persons was performed independently by comparison with a color table, as in the interpretation of microbial tests. Statistical analysis. To evaluate the in uence of the milking order and milk yield on the responses to the ve screening tests studied, a logistic regression model (2) with the stepwise option from the logistic procedure of SAS was employed (32). Variables were analyzed using the following logistic model.

J. Food Prot., Vol. 66, No. 11 ANTIBIOTIC RESIDUE PERSISTENCE IN EWE S MILK 2099 TABLE 2. Summary of the logistic regression model coef cients for penicillin depletion in ewe s milk Assay b 01 a b 02 b b 1 C (%) a BRT MRL CH ATK P&S Delvotest SP Eclipse 100ov Penzym 100 4.8063 4.8550 6.6867 10.2911 3.2016 7.5175 6.0370 9.3303 4.7733 21.1829 21.1549 21.7635 22.7418 20.9108 93.2 93.8 96.2 98.5 90.3 a b 01 5 b 0 was used to estimate the frequency of positive/doubtful 1 negative cases. b b 02 5 b 0 was used to estimate the frequency of positive 1 doubtful/negative cases. c C, concordance coef cients. L ijk 5 logit[p ijk ] 5 b 0 1 b 1 [MO] i 1 b 2 [P] j 1 «ijk L ijk is the logit model, [P ijk ] is the probability for the response category (positive/doubtful 1 negative or doubtful 1 positive/ negative), b 0 is the intercept, b 1 and b 2 are the estimated parameters for the model, [MO] i is the effect of milking order (1 through 7, and 12), [P] j is the effect of milk production, and «ijk is the residual error. The concordance coef cient was applied as a rank correlation between the observed responses and predicted probabilities (32). RESULTS AND DISCUSSION Benzyl penicillin procaine treatment. Table 2 presents coef cients from the logistic regression model. For all assayed methods, the effect of milking order was signi cant (P, 0.0001) but not milk yield. Molina et al. (26) observed that milk production didn t contribute signi cantly to persistence of penicillin in milk from ewes treated intramuscularly. Also, Seymour et al. (34) did not nd signi cant effects for this factor in dairy cows. Nevertheless, sheep with lower milk production took longer to eliminate antibiotics other than penicillin supplied via other administration routes, such as intravenous or intramammary (26). Also, at the end of lactation, when milk production is low, a signi cant effect of milk yield, even by intramuscular route, has been found (5). Higher values for b 1 coef cients showed by Eclipse 100ov (22.7418; Table 2) and Delvotest SP tests (21.7635; Table 2) revealed a steeper slope in the evolution of the positive results for these methods than did coef cients for the other methods (Fig. 1). BRT MRL and Delvotest SP methods had elevated frequencies for doubtful tests in the period between the fourth and seventh milkings (Fig. 1). This result is associated with the b 02 2 b 01 parameter from the logistic model because BRT MRL (7.5175 2 4.8063) and Delvotest SP (9.3303 2 6.6867) methods showed higher b 02 2 b 01 values. For this antibiotic, no doubtful responses were provided with Eclipse 100ov. At the rst milking after the withdrawal period ( ve milkings), positive responses ranged from 2% (Delvotest SP) to 11% (CH ATK P&S), with only the Eclipse 100ov test providing all negative results (Fig. 1), perhaps because the rst methods possess detection limits under the MRLs (between 1.4 and 3 mg/kg for Delvotest SP (7) and Penzym FIGURE 1. Relative frequency of positive (v) and doubtful (h) cases in ewe s milk treated intramuscularly with benzyl penicillin procaine using BRT MRL, CH ATK P&S, Delvotest SP, Eclipse 100ov, and Penzym 100 tests. 100 (3), respectively). However, Eclipse 100ov has a detection limit of 5 mg/kg (6), over the MRLs for penicillin (4 mg/kg (10, 12)). Eclipse 100ov is a speci c ewe s milk test (14), whereas the other tests were developed for cow s milk and subsequently evaluated to apply to ewe s milk (3, 4, 7, 27). On the other hand, there could be a slight variation between the sensitivity when detecting residues in milk forti ed in vitro and milk from treated animals, where natural inhibitors could also interfere. Moreover, small differences in the molecular structure of penicillin, as in penicillin G versus benzyl penicillin procaine, might provoke variations in method responses, as with Delvotest SP. Ampicillin treatment. The results from the logistic model used to estimate the responses for the screening methods are summarized in Table 3. As in the penicillin treatment, only the milking order for the ve tests was signi cant (P, 0.0001). For that reason, the effect of milking order on the frequency of positive and doubtful responses to the assayed methods is shown in Figure 2. Eclipse 100ov (23.3116), BRT MRL (22.4580), and Delvotest SP (22.0065) tests presented the highest b 1 coef cients. This reveals a faster decline of positive responses with these tests throughout the experimental period. As with the benzyl penicillin procaine treatment, the b 02 2 b 01 parameter was higher in BRT MRL (16.6004 2 10.0290) and Delvotest SP (11.2768 2 7.7846) tests, corresponding to the higher frequencies of doubtful tests observed in Figure 2 for BRT MRL and Delvotest SP. These high frequencies could be due to the coincidence of the

2100 BERRUGA ET AL. J. Food Prot., Vol. 66, No. 11 TABLE 3. Summary of the logistic regression model coef cients for ampicillin depletion in ewe s milk Assay b 01 a b 02 b b 1 C (%) c TABLE 4. Summary of the logistic regression model coef cients for amoxicillin depletion in ewe s milk Assay b 01 a b 02 b b 1 C (%) c BRT MRL CH ATK P&S Delvotest SP Eclipse 100ov Penzym 100 10.6290 8.5030 7.7846 13.6978 4.2916 16.6004 10.0424 11.2768 14.4512 5.7963 22.4580 21.7080 22.0065 23.3116 20.9826 94.7 96.5 95.0 98.4 91.5 BRT MRL CH ATK P&S Delvotest SP Eclipse 100ov Penzym 100 6.8887 6.9055 6.5344 9.2268 3.9108 7.9490 7.8484 8.8514 11.6771 5.8315 20.9742 20.8418 21.2277 21.8235 20.6905 92.3 90.1 94.2 96.6 86.7 a b 01 5 b 0 was used to estimate the frequency of positive/doubtful 1 negative cases. b b 02 5 b 0 was used to estimate the frequency of positive 1 doubtful/negative cases. c C, concordance coef cients. a b 01 5 b 0 was used to estimate the frequency of positive/doubtful 1 negative cases. b b 02 5 b 0 was used to estimate the frequency of positive 1 doubtful/negative cases. c C, concordance coef cients. detection limits of the assayed test with the MRLs at the end of the withdrawal time. On the other hand, some authors (22) observed a diminution in the frequency of doubtful cases when analyses were con rmed 24 h after heating to 828C. For this reason, we recommend that laboratories and dairy industries repeat analyses of doubtful results after 24 h to de ne milk use and avoid the entrance of contaminated milk into the food chain. The ampicillin manufacturer s directions establish six milkings as the recommended withdrawal period. As Figure 2 indicates, there are almost no positive responses in the seventh milking, except for the Penzym 100 test, which showed 7%. Sensitivity to the tested methods is close to the ampicillin MRL (4 mg/kg (10, 12)): within 3 mg/kg for Delvotest SP (7) and BRT MRL (15) and 4 mg/kg for Eclipse 100ov (28), Penzym 100 (3), and CH ATK P&S (13). Amoxicillin treatment. Statistical coef cients, obtained by means of the logistic model, to the response of the detection inhibitor methods in milk samples from amoxicillin-treated ewes are given in Table 4. As in previous treatments, Eclipse 100ov (21.8235) and Delvotest SP (21.2277) tests showed the highest b 1 coef cients. In this case, the methods that showed higher doubtful responses were Eclipse 100ov, Delvotest SP, and Penzym 100 (Table 4 and Fig. 3). High frequencies of positive responses were observed (73.4% CH ATK P&S, 51.7% BRT MRL, 28.4% Penzym 100, and 11.3% Delvotest SP) beyond the rec- FIGURE 2. Relative frequency of positive (v) and doubtful (h) cases in ewe s milk treated intramuscularly with ampicillin using BRT MRL, CH ATK P&S, Delvotest SP, Eclipse 100ov, and Penzym 100 tests. FIGURE 3. Relative frequency of positive (v) and doubtful (h) cases in ewe s milk treated intramuscularly with amoxicillin using BRT MRL, CH ATK P&S, Delvotest SP, Eclipse 100ov, and Penzym 100 tests.

J. Food Prot., Vol. 66, No. 11 ANTIBIOTIC RESIDUE PERSISTENCE IN EWE S MILK 2101 ommended withdrawal period of six milkings (Fig. 3), although this frequency was low in the Eclipse 100ov test (2.8%). These observations could be justi ed with the detection limit for each method. In fact, detection limits for the BRT MRL and CH ATK P&S tests are 2 to 3 mg/kg (15) and 3.5 mg/kg (13), respectively, which is why it is possible to detect residues after the withdrawal time. For Delvotest SP (7) and Penzym 100 (3), detection limits are 5 mg/kg, which is near the MRLs (4 mg/kg (10, 12)), and showed high frequencies of doubtful cases (45.1% for Delvotest SP and 44.6% for Penzym 100) at the end of the withdrawal time, as in the ampicillin treatment. By contrast, the Eclipse 100ov test has a detection limit value of 7 mg/kg (6), which explains the lowest positive frequencies after the safety period (2.8%). For amoxicillin, positive and doubtful tests have been detected after the withdrawal period even when using methods with detection limits over the MRLs. These ndings imply that extension of the withdrawal time to more than six milkings is recommended in order to carry out the EU normative. A word of warning to the dairy laboratories and industries might be advisable because, even though farmers follow the manufacturer s label instructions, positive and doubtful results could be obtained in ewe s milk from individual animals when samples are analyzed with BRT MRL, Delvotest SP, CH ATK P&S, and Penzym 100 tests, perhaps because the majority of the tests show detection limits under that of the MLR established for the antibiotics studied. Therefore, it is necessary to adapt these tests so that the detection limits are on par with the MRLs established. However, it is necessary to take into account a different approach in the concept of residue limit because a position that considers the potential technological problems derived from residue persistence, and not just safety troubles, would be more desirable. According to Adriany (1), MRLs only represent safety toxicological levels; however, the microbiological effects in the dairy industry and the safety technological levels are not considered in this evaluation. For this reason, it is important to harmonize the analysis concept between industry and reference laboratories to avoid penalties for producers and to avoid technological problems. ACKNOWLEDGMENTS This work was funded by project 191/IA-40 from the Consejeria de Agricultura y Medio Ambiente (Junta de Comunidades de Castilla-La Mancha, Spain). The authors thank the personnel from ITAP (Diputación de Albacete, Albacte, Spain) for their valuable help in the experiment and Professor K. A. Walsh for her assistance with the preparation of this manuscript in English. Also, authors are grateful to AIM-Analytik in Milch Produktions-undVertriebs GmbH (Munich, Germany), Chr Hansen (Hoersholm, Denmark), DSM Food Specialties (Delft, The Netherlands), and ZEU-Inmunotec (Zaragoza, Spain) for their support. REFERENCES 1. Adriany, A. 2002. Test microbiológicos para inhibidores; test de screening o de con rmación, p. 40. In Proceedings of the XIX Reunión G-Temcal Nuevos retos en los sistemas de calidad de producción de leche para el primer decenio del siglo XXI. Palencia, Spain, 25 26 October. 2. Agresti, A. 1990. Categorical data analysis. John Wiley, New York. 3. Althaus, R., M. P. Molina, M. Rodr guez, and N. Fernández. 2001. Detection limits of b-lactam antibiotics in ewe milk by Penzym enzymatic test. J. Food Prot. 64:1844 1847. 4. Althaus, R., P. Molina, M. Rodr guez, and N. Fernández. 2001. Evaluation of the BRT method for detection of b-lactam antibiotics in ewe milk. Milchwissenschaft 56:568 572. 5. Althaus, R. L., P. Molina, H. Soutullo, C. Peris, M. Mart nez, and N. Fernández. 2001. Antibiotic depletion time and screening methods in milk of Manchega dairy ewes. Options Méditerr. 46:141 144. 6. Althaus, R. L., A. Montero, M. C. Beltrán, R. Llor a, and M. P. Molina. 2002. Detección de residuos de medicamentos en leche de ovejas (Eclipse 100ov), p. 224 228. In Proceedings in XXVII Jornadas Cient cas de la Sociedad Española de Ovinotecnia y Caprinotecnia, Valencia, Spain, 19 21 September 2002. 7. Althaus, R. L., C. Peris, A. Montero, A. Torres, and P. Molina. 2002. Detection limits of antimicrobial agents in ewe milk by Delvotestt. Milchwissenschaft 57:660 664. 8. Anadón, A., and M. R. Martinez-Larrañaga. 1999. Residues of antimicrobial drugs and feed additives in animal products: regulatory aspects. Livestock Prod. Sci. 59:183 198. 9. Anderson, K. L., W. Moats, J. E. Rushing, and J. M. O Carroll. 1998. Detection of milk antibiotic residues by use of screening tests and liquid chromatography after intramammary administration of amoxicillin or penicillin G in cows with clinical mastitis. Am. J. Vet. Res. 59:1096 1099. 10. Anonymous. 1990. Council Regulation (EEC) 2377/90 laying down a community procedure for the establishment of maximum residue limits of veterinary medicinal products in foodstuffs of animal origin. Off. J. Eur. Commun. L224(26 June 1990):1 8. 11. Anonymous. 1992. Council Regulation (EEC) 92/46/EEC laying down the health rules for the production and placing on the market of raw milk, heat treated milk and milk based product. Off. J. Eur. Commun. L268(16 June 1992):1 20. 12. Anonymous. 1999. Commission Regulation (EEC) 508/99 amending annexes I to IV of Council Regulation (EEC) 2377/90 laying down a community procedure for the establishment of maximum residue limits of veterinary medicinal products in foodstuffs of animal origin. Off. J. Eur. Commun. L60(4 March 1999):16 52. 13. Anonymous. 2001. Copan test (CH ATK) antibiotic test instructions. CHR Hansen, Hoersholm, Denmark. 14. Anonymous. 2002. Eclipse 100: test para la detección de substancias antibacterianas en leche. Technical Report ZEU-Inmunotec, S.L., Zaragoza, Spain. 15. Anonymous. October 2002. Detection sensitivity of the AIM BRTtests towards selected antibiotics in mg/kg (ng/mol; ppb). Available at: http://www.aim-bayern.de/english/information/detectionm.htm. Accessed 28 October 2002. 16. Bishop, J. R., and C. H. White. 1984. Antibiotic residue detection in milka review. J. Food Prot. 47:647 652. 17. Buswell, J. F., and D. M. Barber. 1989. Antibiotic persistence and tolerance in the lactating sheep following a course of intramammary therapy. Br. Vet. J. 145:552 557. 18. European Medicines Evaluation Agency (EMEA). 14 July 1999. Antibiotic resistance in the European Union associated with therapeutic use of veterinary medicines: Report and qualitative risk assessment by the committee for veterinary medicinal products. EMEA/CVMP/ 342/99 Final Report. Available at: http://www.emea.eu.int/pdfs/vet/ regaffair/034299.enc.pdf. Accessed 11 August 2002. 19. Freischem, B. J. 2000. Regulation of residues of antimicrobial veterinary medicinal products in food in the European Union. Microb. Ecol. Health Dis. 12(1 Suppl./10):1 5. 20. Heeschen, W. H., and G. Suhren. 1996. Principles of and practical experiences with an integrated system for the detection of antimicrobials in milk. Milchwissenschaft 51:154 160. 21. International Dairy Federation Bulletin. 1991. Detection and con r- mation of inhibitors in milk and milk products. IDF Bull. 258. International Dairy Federation, Brussels, Belgium.

2102 BERRUGA ET AL. J. Food Prot., Vol. 66, No. 11 22. Kang, J.-H., and F. Kondo. 2001. Occurrence of false-positive results of inhibitor on milk samples using the Delvotest SP assay. J. Food Prot. 64:1211 1215. 23. Laroque, L., and G. A. Neville. 1986. A practical evaluation of the Delvotest P Multi Plate test in screening raw milk for antibiotics. J. Food Prot. 49:868 870. 24. Mäyrä-Mäkinen, A. 1995. Technological signi cance of residues for the dairy industry, p. 136 143. In Proceedings of the symposium on residues of antimicrobial drugs and other inhibitors in milk. International Dairy Federation, Kiel, Germany. 25. Mitchell, J. M., M. W. Grif ths, S. A. McEwen, W. B. McNab, and A. J. Yee. 1998. Antimicrobial drug residues in milk and meat: causes, concerns, prevalence, regulations, tests, and test performance. J. Food Prot. 61:742 756. 26. Molina, A., M. P. Molina, R. L. Althaus, and L. Gallego. 2003. Residue persistence in sheep milk following antibiotic therapy. Vet. J. 165:84 89. Available online at: http://www.idealibrary.com, doi: 10.1016/S1090-0233(02)00173-9. 27. Molina, P., R. L. Althaus, M. A. Zorraquino, A. Molina, and N. Fernández. 2002. Evaluation of Penzymt enzymatic test in ewe s milk. Milchwissenschaft 57:33 35. 28. Molina, P., A. Montero, E. Gomis, R. L. Althaus, and A. Molina. 2001. Evaluación del método microbiológico Eclipse para la detección de inhibidores en leche de oveja. In Proceedings of the XVIII Reunião G-Temcal Segurança nos lactic neos. Control na exploração, Lisboa, Portugal, 28 29 September. 29. Morétain, J. P. 1986. Les médicaments vétérinaires et la qualité du lait: le problème des résidus d antibiotiques. RIA, Technicien du lait 1:10 16. 30. Mourot, D., and S. Loussouarn. 1981. Sensibilité des ferments lactiques aux antibiotiques utilisés en médicine vétérinarie. Rev. Med. Vet. 157:175 177. 31. Oliver, S. P., J. L. Maki, and H. H. Dowlen. 1990. Antibiotic residues in milk following antimicrobial therapy during lactation. J. Food Prot. 53:693 696. 32. SAS. 1998. SAS user guide: statistics version 6.12. Statistical Analysis Systems Institute, Cary, N.C. 33. Schwarz, S., and E. Chaslus-Dancla. 2001. Use of antimicrobials in veterinary medicine and mechanism of resistance. Vet. Res. 32:201 225. 34. Seymour, E. H., G. M. Jones, and M. L. McGilliard. 1988. Persistence of residues in milk following antibiotic treatment of dairy cattle. J. Dairy Sci. 71:2292 2296.